Tipping Points in Climate & Biosphere

Lecture by Prof Tim Lenton on Tipping Points in Climate & Biosphere

I’m going to give you a tour through tipping points in climate and biosphere function.  An issue which we tend to problematize and I I guess as we were in the threats and challenges session and I spent 10 years for training myself to be an ever more effective Horsemen of the apocalypse.

I might do a bit of that, I actually want to flip the narrative around and talk about how we can tip transformative change or at least I hope you spend the rest of the two days but let’s remember that we’re only here you know I’m in a conversation breathing an oxygen-rich atmosphere because there been some changes some fundamental changes in Earth history caused by the biosphere the coevolution of life and the planet and in the case of oxygen in the atmosphere it didn’t rise smoothly we now understand it as having gone through at least two major tipping points in Earth history one of them was called the great oxidation half way through the planet’s history the other one was still trying to understand and is tangled up with the rise of animals 600 million years ago and snowball earth events that we might also consider a an extreme kind of climate tipping point but as we look ahead to return we certainly problematize the situation now we’re tending to see negative tipping points we’re tending to suspect that we’re somewhere over here in a little boat perhaps without a paddle or with a mica panel floating downstream it was thinking.

problems and dangerous tipping points ahead we’ve no idea how close they are and we need to change course before we get to here because there’s a lot of inertia and locking in the system so ideally we’d like some early warning signals of what’s coming before we get to the critical dashed line so yeah one of the ways I’ve tried to think.

positively about tipping points is even if they’re bad tipping points we can get early warning of what’s coming and to illustrate that I just do a little mathematical cartoon model where we we start with a system with two basins of attraction two alternative states subjected to some stochastic noise and eventually it’s going to tip from one state to the other and on its before it does that actually it’s showing some warning signals that a tipping point is covering and that’s it was slowing down in response to those fluctuations it was getting from its environment all the time and I hope you could kind of see that visually and we were measuring that on the bottom as this rising indicator of of autocorrelation in time in the system and wells yes mostly I’m going to talk about how we might use that to forewarn us of bad things that happening if we were imagining we might want to tip positive transformative.

transformative change in ecosystems back to a preferred state we might be able to use the same indicators to know when and how to effectively intervene to tip positive change.

so next a little tour through scales just tell you what you might already know but that you know we’re living in a world of complex systems nested within complex systems and nonlinear dynamics and so sometimes not all the time but sometimes you know systems are sensitive and they’re so sensitive that they can exhibit this tipping dynamic and that can help that can happen for couple biological physical systems across all sorts of scales so down at the population level that’s kind of

well-known and indeed some experiments done right back in the 80s and repeated more recently show that classic cases of tipping a collapse of a population like this case of photo inhibition it’s called of algae shading over

hemostat show warning signals before the collapse in other words the system recovers ever more slowly from the same perturbation the closer you get to the tipping point we step up scale a little bit ecologists in a room are well aware that there’s a boost more than half a century of work on alternative regimes or stable states and associated tipping points for ecosystems particularly at quantico systems that picture there of a experimental manipulation of a lake on one side of the divider and not on the other showing two alternative states intricated in oligotrophic lake states that dates back to the 70s

more recently Steve carbon two other colleagues have been showing the in experimentally manipulated leg systems where we know that there’s the

possibility of tipping points that indeed and with experimental

manipulation you can convince yourself there early warning signals before the tipping point either to a eutrophic state or for what’s called a trophic cascade in the food web and there’s some work on finding these signals in historical cases before before

historically observed abrupt shifts of Lake ecosystems or of land systems that change state carrying on up and perhaps adding adding debate as we go but there’s a lively discussion about whether we could tip large scale biome change there’s certainly evidence for multiple but say multiple modes of tree cover for the same for a given rainfall level right across the tropics and that’s what these plots are getting at so for a whole range of rainfalls around the tropics of the planet you can find high tree cover States known otherwise as rainforests and low tree cover States known otherwise in savannas now how much of that by modality in the data reflects a true by stability which means a true underlying self amplifying feedback that can separate these two attractive States well that’s the thing that’s under debate particularly as us humans effectively entwined in the system and disturbing it by changing things like the fire regime where fire is one of these key feedbacks that is reinforced if you like by the grassland State and suppressed by by the forest State well suffice to say that some of the nice recent work like this paper by Burt Weitz is showing that indeed if you factor humans in the equation some of the BI modalities is to do with the correlation of human interference with the state but there’s still some true bi-stability

certainly in the amazon region it may be less spatially extensive and there’s still an open question over what the true mechanisms of a feedback here are that separate the attractors as we would say of course my pet thing has been an even larger scale where something like the prospect of alternative states and tipping points of the amazon is just one of several candidate tipping elements in the climate system – right – defined as being kind of at least sub-continental in scale some of them quite physical to do with the purely physical nonlinear dynamics of say the Atlantic’s meridian or overturning circulation or all the feedbacks involved in losing an ice sheet but other than others being very very biological including the Amazon and their coupling of Sahel vegetation – to the atmospheric circulation of the monsoon regime when we came at this map from like three different sources evidence that these subsystems had shown abrupt shifts in the past and or evidence from models that that we saw about changes in the future in complex models or thirdly that fundamental understanding of the processes involved with give us reason to believe that the self amplifying feedbacks in some of these systems were so strong that they could give this tipping behaviour well ten years on and lots of work has been done trying to work out how close to these candidate climate tipping points actually of this work was ten years ago it was an expert elicitation where the experts were given three different temperature

corridors into the future that’s less than two degrees that’s two to four degrees centigrade sorry I’m British and that’s 48 degrees high-end warming and these black bars are individual experts imprecise probability estimates that likelihood of tipping five different tipping elements maybe we should just say concentrate on the Amazon rainforest dieback for this talk and predictably as the temperature goes up the experts assessment of the likelihood of tipping goes up I think the key take-home message from that was whilst economists on the climate problem had assumed that tipping points were high impact but very low probability events this expert elicitation even if you’re taking the most conservative assumption says that if we carry on business as usual the lower brown probability on the

likelihood of one of these five tipping points is already 56 percent it’s more likely than not so it’s high impact high probability event and in mid warming it’s the lower bound probabilities around 16% with the most conservative assumptions so it’s still a relatively high probability event more recently some colleagues have been digging around in the climate models from the last IPCC report and finding lots of abrupt shifts at sort of regional scales shall we say within the models and the striking result from that was that a load of those abrupt shifts are clustered somewhere between one and a half and two degrees of warming above pre-industrial which I don’t think any of us is quite understood where that’s coming from I’ve got my suspicions of why that’s the case and that’s to do with how we’ve tuned our models in the past to be stable at the present climate but we didn’t push them too far there’s a bias here because we have less scenarios going up to high warming so you have to imagine rewriting the distribution which brings this up as a second peak but still this is a distinct clustering of abrupt shifts that they are lowering forming so yeah it’s a complex coupled system right so you tip one bit of it it probably has some causal debate percussions for other bits and that’s what this map is about again from the expert solicitation and what was slightly worrying was they’re slightly more mathematically positive cause

interactions between tipping events than mathematically negative ones because if we change that language to the emotional language the mathematically positive is the emotionally negative here it’s the possibility that tipping one thing increases the probability of tipping another in another and the dominoes falling I’ll say a bit more about that because the world can’t be so unstable that that’s the norm because we wouldn’t be here to talk about a certificate was but um yeah there’ll be more on that it means that when friends and colleagues put out a picture like this I have a lot of caveats about it as a co-author the idea that that there might be some sort of global scale tipping point I wouldn’t I wouldn’t say that this is accurately described as a tipping point of a hothouse earth but I would say as climate scientists there’s a potential tipping point to a wet house earth where modest levels of warming we know we could commit the irreversible loss of major ice sheets and tens of meters at sea-level rise it will be very difficult to reverse well take a long time to unfold but I think there is some truth in that and in a true risk analysis we can’t rule out possibilities of global scale tipping points given that we’ve seen them in Earth’s history and indeed when you look at the recent ice court recorded Antarctic temperature and co2 over the last 650,000 years if you took the labels off the scales and Game’s that plots to an electrical engineer or maybe a cardiologist and you ask them how stable the system was if they know their subject especially the electrical engineer they’ll tell you this is not a super stable climate system because it’s oscillating between these two bounds it’s doing this sawtooth oscillation it flicks out of the depths of the last ice age incredibly abruptly and it would appear that their self amplifying feedbacks roll the day temporarily before the system is again bounded in the interglacial state so my take on it is the climate system is actually unusually unstable at the time that we come along and we push out to here so far and up to there in my lifetime and deny that so even to a non-scientist audience who would do

that was going to have some serious consequences anyway let’s get back to early warning signals there so let me go back into the ice age to show you that there have been lots of about climate changes during the last ice age 20 out of them called dense guard ocean events including at the end of the Ice Age we zoom in at the end of the Ice Age has one here it’s called the bollocky warming or the following other warming the cool Younger Dryas interval then an abrupt warming into the Holocene the current interglacial and just

cherry-picking now but we don’t we’ve gone back and looked at these past about changes and we’ve looked before them to see if there were early warning signals of what was coming and here I’m going from about the depth of the last ice age through to the end of the Younger Dryas but I’ve cut out the final transition out of the ice age the final abrupt shift there’s a few issues for Shirdi trending the data that there’s a very strong early warning signal of slowing down and the climate dynamics before the end of the last ice age and even if you took out the problems here do trending the data it’s very clear that there are longer lived fluctuations in the climate here than there are back here and that’s the signal that something big was about to happen now we’ve also gone into model worlds to convince ourselves that in our models we can force something bad to happen we can force in this case a collapse of the Atlantic overturning circulation part of the surface

manifestation of that is the so-called Gulf Stream so we forced this model the add fresh water to the North Atlantic until we break the overturning

circulation up there takes a while to switch off then we look at the pattern in the fluctuations before the claps and we see that the system slows down and the fluctuations slow down before the collapse happens and we can get a reliable warning indicator of that and we can get a reliable warning indicator at exactly the latitude where we’ve spent millions of dollars and pounds putting in at monitoring of the

overturning turns out not to be the best place to look in a signal but it works there no ecologists obviously if there’s a climate tipping point there’s going to be bad consequences for ecosystems and this is just going to show you many things but this is if we force that claps of the overturning circulation what happens to the productivity of brass is in our rather crude model of vegetation within the Hadley center model and basically it’s a big hit on net primary productivity in the growing seasons across Europe what I want to turn to you now though is well what if you don’t believe me you know maybe maybe the climate tipping points really are you know still no probability events could we all agree though that we might learn something interesting about how climate variability is changing by applying these same kind of methods of time series analysis and might that have some importance for ecosystem change so what I’ve done here with Chris Bolton is we were just looking at the

observational data at sea surface temperature fluctuations and be very careful here there’s a lot of in filling of the data in the southern hemisphere but it’s better where the ships have been on regular ship tracks in the northern hemisphere they were seeing a pattern where there’s a lot of slowing down at sea surface temperature

fluctuations where it’s actually monitored for a long time across the North Pacific and the North Atlantic which surprised us so he looks a bit more detailed especially at the North Pacific it was a pattern of sea surface temperatures in the North Pacific and how it oscillates around that pattern is a well-known thing or the Pacific decadal oscillation index that’s what it looks like and who knew that if you just run the analysis on the end the NOAA Pacific decadal oscillation index for the last century or so you find out that there’s very robustly rising

autocorrelation of that data set which is another way of saying a slowing down of the fluctuations in in the sea surface temperature

there’s also rising variants which then drops off again more recently why should I care you’re you might be thinking but so what’s that that’s saying that the whole sea surface temperature

variability is becoming slower its fluctuations are more prolonged they’re also overall increasing in amplitude well that’s got a matter for the ecosystem and the default position in most models of the simple models of the marine ecosystem is that it has a time scale probably governed by the zooplankton of around 2 years and then your typical ocean fluctuations probably had a six-month time scale at the start of this in the early 20th century but the ocean fluctuations been slowing down over time so their their timescales coming towards that of the ecosystem which means we see the in a simple model we see the ecosystem become more correlated with the climate variability which means its variability tends to go up even if there’s no increase in variance in the fourth thing the variance in the response will go up because the ecosystem is integrating this slower forcing and I’d like to sort of put it out there that maybe that has something to do is why we’ve got very we have had very well known abrupt shifts in North Pacific ecosystems there was the most famous ones are in the later 20th century those became more likely I would submit because the sea surface temperature variability had slowed down we’ve done some more analysis of the observational record that kind of orangey red areas are showing this increasing variability in observational private this is the models projecting into the future what happens to

temperature variability into monthly in the high end warming scenario look at these hot spots that come out 40 odd percent increase in temperature

variability over the Amazon at the end of the century and about a 20% increase in Southern Africa

big decrease where the Arctic sea ice is gone so it’s no longer making its contributing well the temperature variability

I just focused in on the Amazon is my final example well of course we know we know climate variability is doing crazy things over the Amazon there being a striking droughts in the last decade or so we have some models saying that the Amazon has alternative states not all models but not not just this model Carlos’s mammal also has alternative states in the Hadley center model you can warm the climate warm the climate and you don’t see any response in the forest but you’ve already made a committed change in that model well to a large loss of forest when you’re at 2 or 3 degrees warming you

and he died back because the forest is lagging the forcing it’s lagging the climate we asked whether they were ourselves whether there are early warning signals of that in the model world and there aren’t the conventional ones because the timescale separation isn’t there the forest is being forced very quickly relative to its natural time scale as trees but we can find a system specific early warning which is that they’re basically a given

temperature perturbation causes a larger loss of carbon if it’s a warming or uptake of carbon if it’s a cooling when the baseline temperature goes up and we could use that as a warning signal that things were getting worse for the amazon 48 seconds to go to return to the topic then of putting this into an economic analysis the point out was trying to make in response to Chris’s talk because we’ve done some work where we try to put stochastically uncertain tipping points in the loss of ecosystems into the famous dice model that they’ll Nordhaus just got a nobel prize for and the long and short of it hit here is a model like that has a welfare or utility function it can include unsubstituted all qualities of the biosphere and

ecosystems and non-market values and when you do that you’ve even with very small relatively low risks of a

relatively small loss of non-market benefits that are on substitutable it can radically increase the incentive to mitigate climate change expressed as what levels are set the carbon taxes in the model so let me wrap up 7 seconds to go if business as usual continues then climate and biosphere tipping points are expected to become high impact high probability events early warning systems exist they don’t always work but they can be used an aid to adaptation this isn’t just about climate change it’s about our life support system the ultimate source of all our flourishing and I think that change in the narrative is essential and we should shift to talking about how we could tip

transformative change in the right direction for social and ecological systems.